Electrochemical and PM-IRRAS Characterization of Cholera Toxin

Dec 20, 2012 - Department of Chemistry, Acadia University, Wolfville, Nova Scotia B4P 2R6, Canada. •S Supporting Information. ABSTRACT: A mixed ...
3 downloads 0 Views 3MB Size
Article pubs.acs.org/Langmuir

Electrochemical and PM-IRRAS Characterization of Cholera Toxin Binding at a Model Biological Membrane J. Jay Leitch,† Christa L. Brosseau,†,⊥ Sharon G. Roscoe,∥ Kyrylo Bessonov,‡ John R. Dutcher, and Jacek Lipkowski*,†

§



Department of Chemistry, ‡Department of Molecular and Cellular Biology, and §Department of Physics, University of Guelph, Guelph, Ontario N1G 2W1, Canada ∥ Department of Chemistry, Acadia University, Wolfville, Nova Scotia B4P 2R6, Canada S Supporting Information *

ABSTRACT: A mixed phospholipid-cholestrol bilayer, with cholera toxin B (CTB) units attached to the monosialotetrahexosylganglioside (GM1) binding sites in the distal leaflet, was deposited on a Au(111) electrode surface. Polarization modulation infrared reflection absorption spectroscopy (PMIRRAS) measurements were used to characterize structural and orientational changes in this model biological membrane upon binding CTB and the application of the electrode potential. The data presented in this article show that binding cholera toxin to the membrane leads to an overall increase in the tilt angle of the fatty acid chains; however, the conformation of the bilayer remains relatively constant as indicated by the small decrease in the total number of gauche conformers of acyl tails. In addition, the bound toxin caused a significant decrease in the hydration of the ester group contained within the lipid bilayer. Furthermore, changes in the applied potential had a minimal effect on the overall structure of the membrane. In contrast, our results showed significant voltage-dependent changes in the average orientation of the protein α-helices that may correspond to the voltage-gated opening and closing of the central pore that resides within the B subunit of cholera toxin.



INTRODUCTION Cholera is an ancient and devastating disease that is endemic in many parts of the world, including South America, Africa, and Southeast Asia.1 This disease is caused when people ingest the bacterium Vibrio cholera, which inhabits freshwater lakes, rivers, and estuaries. Once ingested, the bacterium colonizes the small intestine and subsequently secretes a protein toxin.1 The cholera toxin belongs to the AB5 superfamily of toxins, which all contain an active (A) unit and a binding (B) pentameric unit.2 The five subunits within the pentamer bind specifically to monosialotetrahexosylganglioside (GM1), which is present in the outer leaflet of the plasma membrane. Once bound, the toxin is translocated across the plasma membrane, where the A unit undergoes proteolytic cleavage, giving rise to the enzymatically active A1 unit.2 This activated peptide catalyzes the ADP-ribosylation of the α subunit of the heterotrimeric GTP-binding protein (Gs), which renders adenylate cyclase continually active, thereby increasing cyclic AMP levels in the cell.2 Cyclic AMP catalyzes the phosphorylation of chloride channels in the cell, causing them to open and remain open indefinitely. This opening of the chloride channels in the cell membrane causes a massive loss of water and electrolytes into the intestinal lumen, giving rise to the hallmark features of a cholera infection, which include massive diarrhea and dehydration. The disease is generally fatal within 24 h if not properly treated. © 2012 American Chemical Society

The biochemistry of cholera toxin infection is well understood. However, the mechanism by which the toxin is able to cross the plasma membrane remains elusive, as does the extent to which the toxin is transported. It is known that membrane transport of the toxin is dependent on cholesterol being present, and it involves lipid rafts that are GM1-enriched membrane domains.3 A better understanding of how the binding of cholera toxin affects the structures of both the membrane and the toxin itself is crucial to gaining insight into this infection as well as the protein−membrane interactions in general. Early ellipsometry and flow cytometry work has confirmed that the interaction between the binding unit of cholera toxin (CTB) and the lipid receptor GM1 is specific and multivalent in character.4,5 In fact, this specific interaction is often used as a marker of lipid rafts in membrane systems.6 In the absence of GM1, the B unit of cholera toxin is believed to have little specific interaction with pure phospholipid membranes, which was confirmed by AFM studies.7 Forstner et al. used fluorescence correlation spectroscopy and ATRFTIR to characterize changes in the lipid lateral diffusion and membrane phase structure as a result of CTB binding to membrane-incorporated GM1.8 They observed that significant changes in the long-range diffusion of fluorescent probes Received: December 17, 2012 Published: December 20, 2012 965

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

ultrapure (>18.2 MΩ·cm) water system (Millipore). The threeelectrode glass electrochemical cell was equipped with a single-crystal Au(111) working electrode (WE), which was grown, cut, and polished in our laboratory and a gold wire counter electrode (CE), both of which were flame annealed using a Bunsen burner and quenched with Milli-Q water to ensure that the surface was free from contamination. A saturated calomel electrode (SCE) was used as the reference electrode (RE) in the electrochemical measurements. The PM-IRRAS spectroelectrochemical cell was equipped with a 1 in. BaF2 equilateral prism (Janos Technology), a Pt foil (Alfa Aesar, 99.99%) counter electrode, and a Ag/AgCl reference electrode. All potentials are reported versus the SCE unless otherwise stated. 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC) (Sigma-Aldrich), cholesterol (Sigma-Aldrich), 1,2-dimyristoyl-d54-sn-glycero-3phosphocholine (Sigma-Aldrich), and GM1 (Avanti Polar Lipids) were dissolved in a 95:5 mixture of HPLC-grade chloroform (SigmaAldrich) and methanol (Sigma-Aldrich). Sodium fluoride powder (99%, Sigma-Aldrich) was cleaned in a UVO chamber (Jelight) for 30 min prior to each experiment. Both sodium fluoride electrolyte and cholera toxin B subunit (Sigma-Aldrich) solutions were prepared using Milli-Q or deuterated (Cambridge Isotope Laboratories) water. Bilayer Deposition. A combination of the Langmuir−Blodgett (LB) and Langmuir−Schaefer (LS) techniques was employed to fabricate the phospholipid bilayers on the Au(111) surface using the procedure in ref 18 and described in detail in the Supporting Information. The first monolayer deposited onto the Au(111) surface contained only DMPC and cholesterol in a 70:30 mol ratio. The headgroups of DMPC molecules in the adsorbed film faced the metal electrode surface, and the hydrocarbon tails were directed toward the air.18 The second monolayer deposited onto the Au(111) surface modeled the outer (extracellular) leaflet of a cell plasma membrane and contained DMPC, cholesterol, and GM1 in a 60:30:10 mol % ratio as well as the bound cholera toxin binding (CTB) unit. The average content of GM1 in the epithelial cell of human intestine is much lower, on the order of 1%.19 However, it is now well established that GM1 segregates into “rafts” where its concentration is enriched several-fold.20,21 Hence, the content of GM1 used in this study is likely to be close to that of GM1 in these rafts. To make efficient use of the costly toxin protein, a method proposed in ref 22 was modified and used to expose the protein to the biomembrane. Following this incubation period, the monolayer-covered Au(111) electrode horizontally touched the surface of the compressed monolayer and was transferred using Langmuir−Schaefer deposition.18 This model biological membrane complete with bound toxin was then used to perform electrochemical and PM-IRRAS measurements. Electrochemical Measurements. The electrochemical measurements were carried out in an all-glass three-electrode cell using the hanging meniscus configuration.23 The cleanliness of the 0.1 M NaF electrolyte was checked by cyclic voltammetry (CV) and differential capacitance (DC) measurements using experimental procedures described in our previous work.24 Chronocoulometry measurements were performed to determine the charge density at the electrode surface using the procedure described in the Supporting Information. Spectra Collection and Processing. A Nicolet Nexus 870 spectrometer, equipped with an external tabletop optical mount, an MCT-A detector, a photoelastic modulator with a II/ZS50 ZnSe 50 kHz optical head (Hinds Instruments PM-90), and a synchronous sampling demodulator (GWC Instruments) was used to perform the PM-IRRAS experiments. Infrared measurements for each spectral region of interest were optimized by using an appropriate solvent, angle of incidence, and electrolyte thickness between the window and the electrode. The parameters that were used for each region are shown in Table 1 of the Supporting Information. Each thin-layer cavity thickness was determined by fitting the experimental reflectivity spectrum, which is attenuated by the solvent layer between the gold electrode and BaF2 window, to a reflectivity spectrum that is generated from optical constants using the experimental parameters, as described previously.18 A thin-layer cell equipped with planar BaF2 windows (thickness of solution 101.6 μm) was used to determine the transmission spectra of the DMPC/chol/GM1 mixed-vesicle solution

occurred as a result of toxin binding and that these changes were particularly amplified near the lipid gel−fluid transition temperature. ATR-FTIR studies of the methylene stretching vibrations for DMPC/GM1 bilayers demonstrated that toxin binding resulted in an alteration of the fraction of lipids present in the gel phase, suggesting that an ordering of the lipids beneath the binding site is taking place. Miller et al. have conducted numerous studies on mixed GM1/phospholipid monolayers aimed at understanding the nature of the interaction between CTB and GM1 and the resulting effect on membrane structure.9,10 Their recent grazing incidence Xray diffraction studies revealed the coexistence of two phases in the monolayer after toxin binding at a DPPC monolayer containing GM1, and in particular, the formation of structurally distinct and less ordered domains was observed in gel phases upon toxin binding.11 CTB bonded to planar lipid bilayers is known to behave as a voltage-gated ion channel that preferentially transports anions.12 Fluorescence resonance energy transfer (FRET) measurements of CTB-GM1 complexes in vesicles suggest that the orientation of the α-helices of the protein may significantly change with pH and those changes may correspond to the opening or closing of the channel.13 In addition, this study suggests that membrane translocation of the A unit of the protein toxin may be aided by the opening of the pore in the B unit. The specificity of CTB-GM1 binding has been explored for the development of biosensors using monolayers with GM1 self-assembled at a gold electrode surface.14 These sensors have found application in the screening of drugs that prevent cholera toxin binding to the cell membrane and hence prevent infection.15 In the literature, structural studies concerning the binding of CTB to a model membrane were performed using either a single bilayer supported by a Si ATR element8 or monolayers spread at the air−solution interface.9−11 In this work, we investigated the binding of CTB to a single bilayer supported by a gold electrode surface. This system offers unique opportunity to study the potential-controlled interactions of a protein with a model membrane as demonstrated recently by Brand and co-workers.16,17 The purpose of the current study was to probe the binding of cholera toxin to a gold-supported mixed phospholipid (1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC))/cholesterol bilayer with GM1 receptors present in the distal leaflet using electrochemistry and PM-IRRAS in order to answer the following questions: (i) What structural changes are induced in the membrane as a result of the binding of the toxin? (ii) What changes (conformational and orientational) occur in both the phospholipids and toxin after they bind to the membrane? (iii) How does a potential applied to the membrane affect the orientation and conformation of the protein and the lipids? In particular, the ability to monitor this system in situ under potential control allows for an accurate simulation of electric fields typically experienced by real cell membranes. The results presented in this work demonstrate that the orientation and conformation of the lipids show little change as a function of the applied potential. However, the opening and closing of the central pore within the CTB subunit appear to be voltage-dependent and occur via the α-helical components, which are located in the center of the pentamer.



EXPERIMENTAL SECTION

Experimental Setup, Reagents, and Solution Preparation. All glassware was cleaned in a hot mixed acid bath (1:3 HNO3/H2SO4) for approximately 45 min and thoroughly rinsed using a Milli-Q 966

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

with bound cholera toxin. The transmission spectra were used to calculate optical constants for various spectral regions. The calculations and the optical constants are presented in the Supporting Information (Figures SI1−SI4). The optical constants and the experimental conditions reported in Table 1 of the Supporting Information were then used to calculate the spectra of randomly oriented DMPC and CTB molecules in the membrane supported by the gold electrode surface.25 The demodulation technique developed in Corn’s laboratory26,27 was used in this work. A modified version of a method described by Buffeteau et al.28 was used to correct the intensities of the average and difference signals for the PEM response and the difference in the optical throughputs for p- and s-polarized light. Finally, the measured spectra had to be background corrected for the absorption of IR photons in the thin-layer cavity. The background-corrected spectra are plotted as ΔS, which is proportional to the absorbance A of the adsorbed molecules as a function of wavenumber. The relationship for ΔS is shown in eq 1

ΔS =

2(Is − Ip) Is + Ip

mixed phospholipid membrane in the absence [curve 2 (red)] and in the presence [curve 4 (blue)] of the bound cholera toxin. These plots show that the bilayer is adsorbed onto the electrode surface between 400 and −100 mV versus SCE in the absence of the toxin; however, this range is from 400 to 100 mV versus SCE after binding CTB to the outer surface of the membrane. At more negative potentials, the charge density curves for the bilayer-covered electrodes progressively decrease to merge with the curve for the pure supporting electrolyte at the most negative potentials (between −1.0 and −1.2 V), indicating electric-field-driven dewetting of the bilayer from the metal surface. Independent neutron reflectivity (NR) experiments on similar systems demonstrated that the film, in this state, is detached from the electrode surface but remains in close proximity to the metal surface separated by a cushion of electrolyte with a thickness of ∼1 nm.33 There is a very dramatic increase in the surface charge density when the toxin is bound to the membrane and a very large negative shift in the potential of zero charge, which suggests that the cholera B toxin has a net negative charge. This phenomenon can be explained by the fact that the isoelectric point for the B subunit ranges between 7.4 and 7.634,35 and the sodium fluoride electrolyte had a pH of ∼8.5. As a result, the exterior of the toxin would carry a net negative charge. To ensure that this large increase in the charge density on the surface was not due to the specific binding of the protein to the gold electrode directly via the sulfur groups on the cysteine residues, an aliquot of cholera toxin B was injected into an electrochemical cell that contained only the supporting electrolyte. The resulting charge density curves showed no sizable increase when compared to that of the bare Au(111) electrode in pure electrolyte, which suggests that the toxin has a very small binding affinity for the bare gold substrate. To investigate the degree of nonspecific toxin interaction within the membrane, a bilayer was prepared in the absence of GM1 in the outer leaflet. The resulting charge density from this experiment is represented by curve 3 (green) in Figure 1. An increase in the overall charge density can be observed for the range of potentials between −450 and 400 mV when compared to that in the absence of cholera toxin, and both of these charge density curves merge at the potential of zero charge. These data suggest that cholera toxin has some affinity for biomembranes that lack the receptor; however, the overall magnitude of this interaction is still much smaller than that for the GM1/CTB membrane systems. This was confirmed by recording PMIRRAS spectra for the membrane without GM1 as shown in Figure SI5 of the Supporting Information. PM-IRRAS of the Membrane Lipids. Acyl Chain Region. The C−H stretching region, which is located between 2800 and 3000 cm−1, provides valuable information regarding the conformation and orientation of acyl chains of DMPC in the membrane. Because CTB also contains CH2 groups whose vibrations overlap with those of the membrane, a phospholipid with deuterium-substituted acyl chains (DMPC-d54) was used to separate the IR signatures of the CH2 vibrations of the bilayer from those of the toxin. Therefore, the measurements of C−D stretching were used to characterize the membrane. Figure 2a shows the PM-IRRAS spectra of the C−D stretching region of DMPC-d54 centered at 2100 cm−1 for selected potentials. The top curve in Figure 2a is the calculated spectrum of randomly oriented molecules determined from optical constants obtained from the transmission spectra of DMPC-d54 + cholesterol + GM1 vesicles in H2O. The other

≈ 2.3Γε = 2.3A (1)

where Γ is the surface concentration of the adsorbed species and ε is the decimal molar absorption coefficient of the adsorbed species. The error analysis in the PM-IRRAS measurements and data processing has been described in ref 29. Helix Tilt-Angle Measurements. The cholera toxin B pentamer structure30,31 (PDB ID 3CHB) was obtained from the RCSB Protein Data Bank. The structure was aligned along the z axis using the editconf program in the GROMACS 4.5.5 package.32 The helix tilt angle with respect to the central z axis passing through the geometrical center of the CTB structure was computed with the help of the g_bundle utility from the GROMACS 4.5.5 package. The g_bundle defines the helix axis using the backbone Cα atoms of each helix residue followed by angle calculation between the helix and z axes. Further details of this procedure are described in the Supporting Information.



RESULTS Electrochemistry. The potential-dependent behavior of CTB bound to the bilayer was initially characterized with the help of charge density curves. The charge density curves were used to determine the potential range where the membrane remains stable at the electrode surface and to examine the overall effect of nonspecifically adsorbed proteins. Figure 1 shows the surface charge density curves for the bare Au(111) electrode [curve 1 (black)] and the electrode covered with the

Figure 1. Charge density curves for bilayers composed of DMPC/ Chol/GM1 (2), DMPC/Chol/CTB (3), and DMPC/Chol/GM1/ CTB (4) bilayers supported on a Au(111) electrode surface. The black curve (1) represents the charge density for a bare electrode in a 0.1 M NaF supporting electrolyte. 967

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

D spectral region is composed of six individual bands where the two bands located at 2094 and 2195 cm−1 correspond to the asymmetric, νas(CD2), and symmetric, νs(CD2), methylene stretching vibrations. The positions of these bands provide unique information about the packing and conformation of the phospholipid tails within the membrane, and the intensities of the bands are used to calculate the average tilt angle of the acyl chains. The positions of the band maxima for the νas(CD2) and νs(CD2) vibrations are plotted as a function of potential in Figure 3. The blue filled symbols plot data in the presence of

Figure 3. Peak positions of the vas(CD2) (■, □) and vs(CD2) (▼,▽) bands for the three-component membrane in the absence (open red symbols) and in the presence (solid blue symbols) of bound toxin.

CTB, and the red open symbols plot the data for the bilayer in the absence of CTB, taken from our previous study.36 For the all-trans conformation of the acyl chains, the νas(CD2) and νs(CD2) band positions are 2191 and 2088 cm−1, respectively.37 The frequencies of the CD2 bands shown in Figure 3 are several wavenumbers higher than these numbers, indicating that acyl chains are partially melted and contain a certain percentage of gauche conformations. This behavior is consistent with the presence of cholesterol in the membrane.37 The attachment of CTB has a negligible effect on the frequency of the νas(CD2) band. However, the frequencies of the νs(CD2) vibrations are slightly lower in the presence of CTB. This behavior suggests that, upon the binding of CTB, the acyl chains become slightly more ordered, in agreement with the properties of the DMPC/GM1 single bilayer supported on Si as reported in ref 8. However, in the present case this effect is less pronounced because of the presence of cholesterol. The data in Figure 3 also show that the band positions display a weak red shift at positive potentials. This shift is more pronounced for the νs(CD2) band, which indicates that, with increasing potential, the chains become slightly more ordered and the bilayer is progressing toward a more gel-like state. To quantify the orientation of the DMPC/Chol/GM1 membrane, the integrated intensities of the deconvoluted IR bands were determined for the νas(CD2) and νs(CD2) bands. The integrated intensity of an IR band depends on the angle between the directions of the transition dipole moment of a given vibration and the electric field of the photon. The electric field vector of the p-polarized photon is perpendicular to the metal surface. Therefore, the angle (θ) between the direction of

Figure 2. (A) PM-IRRAS spectra for the C−D stretching region of the DMPC alkyl chains for selected potentials versus SCE. The spectral deconvolutions of the average experimental film and the calculated randomly oriented film are shown in B and C, respectively.

five curves below, plot the PM-IRRAS spectra acquired for the selected electrode potentials indicated in the figure, which show that the band intensity changes as a function of potential. To quantify the IR results, the spectra were deconvoluted into individual component bands. The spectral deconvolutions of the C−D stretching region are presented in Figure 2B,C for the average experimental data and for a randomly oriented film calculated from the optical constants, respectively. The deconvolutions of the spectra for individual potentials are shown in the Supporting Information (Figures SI6). The Fourier self-deconvolution of the IR spectra (described in Figure SI7 of the Supporting Information) revealed that the C− 968

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

monolayer and the increased tilt of the lipid chains, caused by the binding of the toxin. Phosphate and Glycerol Stretching Regions. The bands that originate from the asymmetric phosphate group of the lipids are located in the range of frequencies from ∼1300 to 1200 cm−1. The intensities of these bands are weak and show a minute dependence on the applied potential. To improve the S/N, the bands acquired for several potentials were averaged and the averaged spectrum is plotted in Figure 5. For

the transition dipole and the surface normal can be determined using the following equation38 1 A(E) cos2 θ = 3 A(random) (2) where A(E) is the integrated intensity of the band at a given potential and A(random) is the integrated intensity of a band in a monolayer of randomly oriented molecules. The A(random) value was calculated from the optical constants, reported in the Supporting Information, using the experimental conditions listed in Table 1 of the Supporting Information according the procedure described in ref 25. The calculated angles between the surface normal and the transition dipole moments of the νs(CD2) and νas(CH2) bands of DMPC-d54 in the three-component membrane are shown in Figure SI8 of the Supporting Information. The angles of the transition dipole moments for both the νs(CD2) (θvs) and νas(CH2) (θvas) vibrations show no dependence on the applied electrode potential and are equal within the limits of experimental uncertainties. The direction of the transition dipole of the methylene symmetric stretch aligns along the bisector of the CD2 plane, and the direction of the transition dipole of the asymmetric stretch lies along the line joining the two hydrogen atoms of the CD2 group. Both vectors are positioned within the plane of the methylene group, which is perpendicular to the line of the trans fragment of the hydrocarbon chain. Therefore, it is possible to calculate the average tilt of the trans fragments of the deuterated chains of DMPC (θchain) within the membrane from the experimental θvas and θvs values using the following expression:39 cos2 θvas + cos2 θνs + cos2 θchain = 1

(3)

Figure 5. (A) Qualitative comparison of the average νas(PO2−) and C−O−C stretching regions for the three-component bilayer in the absence (red) and presence (blue) of the cholera toxin. (B) Model of a planar C−C(O)−O−C moiety.

The average tilt angle of the trans fragments of acyl chains is plotted as a function of potential in Figure 4. For comparison,

comparison, the average spectrum for the bilayer without CTB is also included in Figure 5. These spectra are dominated by a broad νas(PO2−) band centered at ∼1230 cm−1. This band is weakly affected by the bonding of CTB, indicating that the attachment of the toxin has little effect on the hydration and orientation of the phosphate group. ATR studies on oriented phospholipid bilayers revealed that the νas(PO2−) band is extremely sensitive to hydration, shifting in frequency from 1250 cm−1 for a dry lipid bilayer to 1230 cm−1 for a hydrated lipid bilayer.40,41 Therefore, the frequency of the νas(PO2−) band indicates that the phosphate group is hydrated in the two bilayers. In the spectra of Figure 5, the band at ∼1175 cm−1 is due to the νas(C−O−C) vibration of the glycerol ester group. This band is more intense in the absence of bound toxin,36 indicating that the orientation of the C−O−C moiety changes in the presence of the protein. According to Fringeli,40,41 this band can be assigned to the C−O−C stretch of the γ chain and the direction of the transition dipole of this band is parallel to the chain. Therefore, the decrease in the C−O−C band intensity in the presence of cholera toxin correlates well with the increased tilt of the chains observed in Figure 4 from the CD2 stretching vibration data.

Figure 4. Tilt angle of the trans fragment of the alkyl chains as a function of electrode potential for a three-component membrane in the absence (□) and presence (■) of bound toxin.

the tilt angles of the trans fragments in the membrane without CTB, taken from a previous study,36 are also included in this figure. The results show that the attachment of CTB causes an increase in the tilt angle by about 10°. This result is consistent with recent X-ray diffraction studies of CTB binding to a phospholipid monolayer at the air−solution interface by Miller et al.11 that show the degradation of the long-range order in the 969

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

out-of-phase components are separated by 5 cm−1. In addition, the position of the ν(CO) band also depends on the hydration state of the ester group. In hydrated phospholipids, this band splits into two components with maxima at ∼1744 and ∼1725 cm−1, corresponding to the non-hydrogen-bonded and hydrogen-bonded ν(CO), respectively.44,45 Because of the complex nature of this band in DMPC, it was not deconvoluted into the hydrogen-bonded and non-hydrogenbonded components in this study. However, the overall shape and position of the CO vibrations depend on the hydration of the ester group because the band position shifts to lower frequencies when the ester group becomes more hydrated. Therefore, the position of the global maximum of this band may be used to assess the degree of hydration qualitatively. To monitor this relative change in membrane hydration, the location of ν(CO) band maxima in the PM-IRRAS spectra is plotted as a function of electrode potential in Figure 6B. The shape and intensity of the ν(CO) band demonstrate little change with the electrode potential; consequently, the band positions vary in a narrow range between 1738 and 1740 cm−1. For comparison, the ν(CO) band maxima determined earlier for the bilayer without CTB36 are also included in Figure 6B. These values are significantly lower, suggesting that the attachment of CTB restricts the hydration of the ester group in the supported bilayer. However, the frequencies of the ν(CO) band maxima for the CTB bound supported bilayer are 5 to 7 cm−1 lower than the spectrum obtained for a suspension of vesicles, indicating that the CO group is more hydrated when the bilayer is supported on the metal surface. Apparently, the binding of CTB to the outer surface of phospholipid vesicles further restricts the hydration of the CO moiety. Figure 6C plots the angle between the surface normal and the transition dipole moment of the CO vibration for the bilayer in the absence and presence of CTB. Because of its complexity and the significant changes in the hydration state of the CO bond, these data points are considered to be approximate values. The direction of the transition dipole of this band is approximately normal to the acyl chains.42 Therefore, the large angles observed for the CO stretching are consistent with the smaller tilt angle values of the acyl chains. In addition, the CO dipole angles in the presence of the toxin are smaller and display a weak dependence on the applied electrode potential, which is consistent with the larger tilt angles that were observed for the acyl chain region, as previously discussed. Therefore, all data obtained from the different spectral regions are in very good agreement In summary, PM-IRRAS studies of the membrane lipids demonstrated that the binding of CTB to the membrane leads to an increase in the tilt of the chains but has a small effect on the chain conformation. The attached CTB shields the glycerol moiety from the water molecules, and as a result, the carbonyl groups are less hydrated. In contrast, its presence does not affect the hydration of the phosphate group. This is not a contradiction. The carbonyl group is electroneutral and is located at the boundary between polar and nonpolar regions of the bilayer. In contrast, the phosphate group is negatively charged and is located in the polar region of the bilayer. The properties of the bilayer with attached CTB determined in this work are in general agreement with the properties observed for this system using different models (single bilayer at supported at silicon8 or a monolayer spread at the air−solution interface9−11). Significantly, our results demonstrate that the

Figure 6A shows spectra for the ester carbonyl group stretching vibration. The top thick line shows the spectrum

Figure 6. (A) PM-IRRAS spectra for the carbonyl stretching region of the DMPC phospholipids for selected potentials versus SCE and (B, C) the corresponding peak maxima for the CO stretching vibration in the absence (□) and presence (■) of the cholera toxin B subunit.

calculated from optical constants determined for a suspension of vesicles. The five lower curves plot the PM-IRRAS spectra for the bilayer supported by the gold electrode at various electrode potentials. The ester CO band located at ∼1730 cm−1 is quite complex.42 DMPC molecules have two conformers, DMPC-A (80%) and DMPC-B (20%).43 The normal coordinate simulation by Bin et al.42 shows that vibrations of the ν(CO) band in the β and γ acyl chains are coupled, causing the band to split into two bands corresponding to the in-phase and out-of-phase motions of the atoms. In the case of the predominant DMPC-A conformation, the coupling is weak, and the in-phase and 970

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

Information Figure SI2) using a thickness value of 32 Å,10,11 are shown in Figure 7. The PM-IRRAS spectra are in good

properties of the membrane change very little with the applied potential. This is important because it shows that the dewetting of the membrane from the metal surface has a minimal effect on the membrane structure. The bilayer supported at a gold electrode surface is a good model for describing potentialcontrolled changes in the orientation of the protein described in the next section. Amide Stretching Region of the Cholera Toxin. The amide bands that arise from the vibrations of the peptide backbone provide useful information on the secondary structure of polypeptides and proteins. An analysis of the peptide group in model compounds and polypeptide systems allows for the assignment of these characteristic bands.46,47 Differences in the hydrogen bonding and dipole−dipole coupling result in changes in the vibrational frequency of the various amide bands, which are characteristic of different elements within the secondary structure of the protein. The amide I−III bands are most frequently used in the infrared analysis of proteins. The amide I band arises primarily from the CO stretching of the amide group47 whereas the amide II and amide III bands are primarily due to N−H bending, with a contribution from C−N stretching vibrations.47 Both theoretical and experimental observations of model peptides indicate that there is a good correlation between the amide I band frequency and the presence of specific secondary structures.46,47 This correlation was further supported by experiments on a vast range of soluble proteins undertaken in different laboratories.47−49 In a D2O electrolyte, bands in the spectral range of 1620−1640 and ∼1680 cm−1 are attributed to a β-sheet structure. Polypeptide fragments in random orientation (random coil or turn) are localized around the 1644 cm−1 region (1640−1648 cm−1). Proteins and polypeptides that assume α-helical conformations give rise to infrared bands in the spectral range of 1650−1658 cm−1.46−48 Typically, the orientational analysis of protein secondary structure involves the determination of the dichroic ratio of the peak heights or integrated areas of the amide I and II bands.49−54 Although this analysis procedure is extensively used, it suffers from several disadvantages. First, the ratio of the amide I and II bands must be performed on dried samples because both H2O and D2O will complicate analysis. In addition, the amide II band cannot be deconvoluted into different secondary structure elements, limiting this approach to peptides or proteins with one predominant structural feature. Furthermore, this analysis is more complicated for α-helical structures because it is now recognized that lengthy α-helices give rise to two distinctively different types of vibrations, A-like modes and E1-like modes. The A-like normal mode refers to the normal mode having all eigenvector elements in phase, and the E1-like mode refers to the normal mode having alternating eigenvector elements, meaning that the eigenvector elements of the local nearest neighbor amide I modes are out of phase with each other.47,54 These different modes arise at different frequencies. For the amide I band, for example, the A-like mode is centered at ∼1651 cm−1 and the E1-like mode is centered at ∼1657 cm−1.47,53 For the amide II band, the frequency location of these modes is more difficult and complicates the analysis. Below is the description of the quantitative analysis of the amide I region that was used to obtain key orientational information in this work. The PM-IRRAS spectra for selected potentials and the spectrum for a randomly oriented CTB film at the gold surface, which was calculated from optical constants (Supporting

Figure 7. PM-IRRAS spectra of the CO stretching region of the amide I band for CTB-bound membranes at selected potentials.

agreement with the spectra of GM1-bonded CTB in vesicles published previously by Surewicz et al.55 The amide I band can be deconvoluted into its component bands, which correspond to the different secondary structural elements of the protein. A Fourier self-deconvolution (FSD) is required to determine the positions of the component bands because the amide I band is broad and featureless. The positions of the peaks determined from FSD are then used for the deconvolution of the broad amide I band contour into its components. Despite the fact that the curve-fitting procedure requires subjective decisions, the prediction of the percentage of each type of secondary structural element in the protein from the area of the component bands is surprisingly accurate.46−48 The bands in PM-IRRAS spectra are about twice as strong as those calculated for the random film, and their intensity gradually increases when the potential is scanned in the negative direction. Because the variable potential constitutes a perturbation, generalized two-dimensional correlation spectroscopy (2D-COS)56 can be used to assist in the identification of the bands present in the PM-IRRAS spectra. In this manner, the confidence level of the resulting deconvolution of the amide I can be further increased. In Figure 8A, the Fourier self-deconvolution was performed on the amide I region of a selected PM-IRRAS spectrum (Figure SI9 in the Supporting Information). It revealed the presence of six sub-bands centered at ∼1618, ∼1627, ∼1655, ∼1667, ∼1673, and ∼1691 cm−1, in good agreement with ref 55. The 2D synchronous correlation analysis of the PM-IRRAS spectra is shown in Figure 8B. The amplitudes in the 2D correlation analysis correspond to the magnitude of intensity deviations from the reference spectrum. The positive peaks located along the diagonal represent the autocorrelation function of spectral intensity variations caused by a change in the applied potential. The diagram shows a very strong autocorrelation peak at (1655, 1655), strong autocorrelations at (1627, 1627), and weak autocorrelation peaks at (1618, 1618), (1679, 1679), and (1691, 1691). The peak at ∼1667 cm−1 resolved by FSD has no autocorrelation peak in the 2D correlation plot. Strong autocorrelation peaks indicate significant changes in these band intensities as a function of applied potential. Conversely, weak autocorrelation peaks or the absence of the autocorrelation peaks indicates a weak dependence or independence of these band intensities on the applied potential. The peaks in off-diagonal positions are the 971

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

with the applied potential. The 2D correlation analysis also reveals a weak cross-correlation peak at (1655, 1642) indicating the presence of a band at 1642 cm−1 that was not resolved by the FSD procedure. The amide I region for the PM-IRRAS and the transmission spectrum of cholera toxin bound to the bilayers were deconvoluted using a mixed Gaussian + Lorentzian fit assuming that seven bands are present in this region. The result of this deconvolution is shown in Figure 8A. The fwhm of the five deconvoluted sub-bands that make up the majority of the amide I band varied from 15 to 19 cm−1. The band at ∼1618 cm−1, assigned to side-chain vibrations according to ref 55, had an average fwhm of 11 cm−1 whereas the average fwhm for the 1691 cm−1 band, which corresponds to β turns according to ref 57, was 11 cm−1. The strong band at ∼1627 cm−1 and the bands at 1673 and 1680 cm−1 belong to β-sheet structures. The bands at 1642 and 1667 cm−1 are due to random structures. Finally, the strong band at 1655 cm−1 belongs to α-helical structures. To relate the assigned bands to the corresponding structural elements, a top view of the “ribbon” model of the protein, taken from the CTB protein database (PDB ID 3CHB),30 is shown in Figure 8C. The average angle of the helices of the CTB crystal structure was 18.7°. The CTB unit is roughly cylindrical with a vertical height of 3.2 nm and a radius of 3.1 nm. The α-helices are tilted, and the unit has a conical shape. The central channel lies along the five-fold axis, and its diameter changes from 1.1 nm at the carboxyl end to a 1.6 nm channel diameter at the amino end. The antiparallel β-sheets surround the central α-helices.30,31 In an effort to obtain information regarding the orientation of the bound protein toxin at the surface, the β-sheet, α-helix, and random IR bands at 1627, 1655, and 1667 cm−1, respectively, were used to calculate the angles between the transition dipole moment and surface normal using eq 2. The calculated dipole angles are plotted as a function of applied potential in Figure 9. These bands were selected because the

Figure 8. (A) FSD (dotted line) and resulting spectral deconvolution of the amide I band showing the contribution from secondary structure for the experimental film at 160 mV. (B) Two-dimensional correlation spectrum for the amide I band as a function of the electrode potential (referenced to 400 mV). (C) Ribbon model of the cholera toxin B subunit taken from the RCSB protein crystal data bank (PDB ID 3CHB). Each of the five central helices are color-coded and numbered with respect to the distance from the N terminus. The tilt angles in degrees are shown in the brackets in the legend. The z axis passing through the center of mass (blue dot) is shown by the black line.30−32

Figure 9. Angles of the transition dipole moments of the β-sheet, αhelix, and turn components of the amide I band for CTB bound membranes as a function of the electrode potential vs SCE.

error in the calculated angle is minimized by the large intensity and low spectral overlap. The angle for the random structures is independent of potential. This suggests that the orientations of the random structures of the cholera toxin B subunit are independent of the electrode potential. In contrast, the angle for the β-sheet structure increases and the angle for the αhelical structure decreases with increasing potential. This result is consistent with the 2D correlation analysis, which showed no

cross-correlation peaks. Positive cross-correlations (marked in red) imply that the two correlated peaks change in the same direction as a function of applied potential. In contrast, negative cross-correlation peaks (marked in blue) at (1655, 1627) and (1655, 1618) indicate that the two bands at 1618 and 1627 cm−1 and the band at 1655 cm−1 change in opposite directions 972

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

autocorrelation band at 1667 cm−1 and the negative sign of the (1655, 1627) cross-correlation band. The agreement between the two types of data analysis indicates that no major errors were made in the band deconvolution procedure. The negative sign of the cross-correlation band (1655, 1627) shows that the potential-induced movements of the α-helices and the β-sheets are interrelated. When this data is combined with the crystal structure of the cholera toxin B subunit, it suggests that the βsheet peptides, which act as a structural support for the αhelices, play a role in the opening and closing of the CTB channel.12 The changes in the angle between the transition dipole of the amide I band and the surface normal are more pronounced for α-helices, indicating a potential-induced orientational change of the helical structure of the toxin. This is important because the central pore of the toxin contains five α-helices and the exact role of this pore remains elusive. The angle (α) between the helical axis and the direction of the transition dipole moment of the amide I vibration was estimated to be between 34 and 38°.53,54,58 This estimate can be used to calculate the average tilt angle of the helices with respect to the surface normal by taking the following approach. First, the order parameter corresponding to the angle between the transition dipole moment of the amide I band, SCO (amide I band is predominantly the vibration of the CO group of the amide bond), was calculated using the expression50 1 SCO = (3 cos2 θCO − 1) (4) 2 Next, the order parameter of the amide I band was used to calculate the order parameter of the helical axis from the following relationship:50,53,54 S helix =

2SCO (3 cos2 α − 1)

(5)

Finally, the average tilt angle (γ) of the helical axes was calculated from its order parameter using the formula 1 S helix = (3 cos2 γ − 1) (6) 2 Figure 10A shows the average tilt angles of the helices as a function of applied potential with two models depicting the open and closed channel configurations of the channel. There are two sets of values that correspond to the upper and lower limits of the angle between the direction of the transition dipole of the amide band and the helix axis used in the calculations. At the positive potentials, the average tilt angle resides between 10 and 18° and increases to a value between 30 and 35° at the negative limit of the potential. To assist the discussion of these changes, Figure 10B shows the plot of the charge density for the CTB bound bilayer at the Au(111) electrode surface. The correlation between the two sets of data is excellent, and it shows that the tilt of the helices is small when the bilayer is directly deposited onto the electrode surface at positive potentials and increases when the bilayer begins to be detached from the metal at negative potentials. The maximum value of the tilt angle is observed at ∼−0.4 V, which correlates with the inflection point on the charge density curve. The limiting value of 30−35° is attained at the most negative potentials where the bilayer is detached from the membrane surface and is suspended on an ∼10-Å-thick cushion of electrolyte.33 This potential-induced reorientation of the helices can occur via (i) a change in the membrane environment at the metal surface as a

Figure 10. Comparison of the (A) tilt angle (γ) of the major axis of the α-helical components of the cholera toxin B subunit using a reference angle of 34° (■) and 38° (◆) in the closed (left) and open (right) channel states, (B) charge density, and (C) tilt angle of the alkyl chains of the DMPC bilayer as a function of the electrode potential.

function of potential or (ii) the opening and closing of the αhelices within the channel CTB protein. To address this issue, the bottom panel in Figure 10C plots the average angle between the trans fragments of the acyl chains of DMPC molecules as a function of the electrode potential. In contrast to the pronounced changes in the tilt of the helices, the tilt angle of the acyl chains shows a weak dependence on the electrode potential, and as a result, one can conclude that the changes in the orientation of the helices are not induced by changes in the membrane structure. Using the crystallographic data, the average tilt of the five α-helices with respect to the axis of the pore was determined to be ∼19°. This number is lower than the tilt with respect to the surface normal at most negative potentials but larger than the tilt with respect to the surface normal at the largest positive potentials. The larger tilt of the five α-helices with respect to the surface normal compared to that with respect to the axis of the pore can be understood because the pore axis does not need to be perfectly parallel to the surface normal. However, tilt angles smaller than 20° 973

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

protein. This material is available free of charge via the Internet at http://pubs.acs.org.

indicate unequivocally that the CTB channel can be opened by the modulated potential because, as the structure in Figure 8C shows, a smaller tilt should result in an overall widening of the CTB pore. The CTB channel should therefore be closed at E < 0.1 V and opened at more positive potentials. The minimum pore radius in the closed position was found to be 1.1 nm.31 Our results suggest that the CTB pore radius could be opened up to at least ∼1.6 nm. This value is close to the 2.1 ± 0.2 nm of the CTB-channel diameter estimated by Krasilnikov et al.12 via conductance measurements. The opening diameter of ∼2.1 nm was considered to be wide enough to allow for the transport of the A subunit of the cholera toxin protein.12 Therefore, the potential-modulated changes in the orientation of the helices are an important result because they provide a possible explanation of the pore-opening phenomenon that is speculated to occur during membrane translocation of the cholera toxin A subunit. In our opinion, this is the most plausible model to explain the data. The imaging experiments by Thompson et al.59 and Mou et al.60 demonstrated that CTB molecules are uniformly distributed at the surface of phosphocholine bilayers with a GM1 content up to 10%. This is an indication of repulsive interactions, and it is unlikely that such interactions will be changed by the applied potential to cause the change in the tilt of the helices. The collective distortion or twist of all five helices, instead of the proposed inclination toward the pore center, may be invoked as well. However, the availability of free space in the central pore makes the proposed inclination the most rational explanation of the present results.



Present Address ⊥

Department of Chemistry, Saint Mary’s University, Halifax, Nova Scotia, Canada B3H 3C3.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was funded by an NSERC research grant. J.L. and J.R.D. acknowledge the Canadian Foundation for Innovation (CFI) for Canada Research Chair Awards. We thank Dr. Jarek Majewski (Los Alamos National Laboratory) for encouraging us to initiate this study.



REFERENCES

(1) Colwell, R.; Anwar, H. Marine Ecosystems and Cholera. Hydrobiologia 2001, 460, 141−145. (2) Torgersen, M.; Skretting, G.; Deurs, B.; Sandvig, K. Internalization of Cholera Toxin by Different Endocytic Mechanisms. J. Cell Sci. 2001, 114, 3737−3747. (3) Lencer, W.; Saslowsky, D. Raft Trafficking of AB5 Subunit Bacterial Toxins. Biochim. Biophys. Acta 2005, 1746, 314−321. (4) Vénien-Bryan, C.; Lenne, P.-F.; Zakri, C.; Renault, A.; Brisson, A.; Legrand, J.-F.; Berge, B. Characterization of the Growth of 2D Protein Crystals on a Lipid Monolayer by Ellipsometry and Rigidity Measurements Coupled to Electron Microscopy. Biophys. J. 1998, 74, 2649−2657. (5) Lauer, S.; Goldstein, B.; Nolan, R. L.; Nolan, J. P. Analysis of Cholera Toxin-Ganglioside Interactions by Flow Cytometry. Biochemistry. 2002, 41, 1742−1751. (6) Kenworthy, A. K.; Petranova, N.; Edidin, M. High-Resolution FRET Microscopy of Cholera Toxin B-Subunit and GPI-Anchored Proteins in Cell Plasma Membranes. Mol. Biol. Cell 2000, 11, 1645− 1655. (7) Yuan, C.; Johnston, L. J. Atomic Force Microscopy Studies of Ganglioside GM1 Domains in Phosphatidylcholine and Phosphatidylcholine/Cholesterol Bilayers. Biophys. J. 2001, 81, 1059−1069. (8) Forstner, M. B.; Yee, C. K.; Parikh, A. N.; Groves, J. T. Lipid Lateral Mobility and Membrane Phase Structure Modulation by Protein Binding. J. Am. Chem. Soc. 2006, 128, 15221−15227. (9) Miller, C. E.; Majewski, J.; Kjær, K.; Weygand, M.; Faller, R.; Satija, S.; Kuhl, T. L. Neutron and X-ray Scattering Studies of Cholera Toxin Interactions with Lipid Monolayers at the Air-Liquid Interface. Colloids Surf., B 2005, 40, 159−163. (10) Miller, C. E.; Majewski, J.; Faller, R.; Satija, S.; Kuhl, T. L. Cholera Toxin Assault on Lipid Monolayers Containing Ganglioside GM1. Biophys. J. 2004, 86, 3700−3708. (11) Miller, C. E.; Majewski, J.; Watkins, E. B.; Kuhl, T. L.; Part, I. An X-ray Scattering Study of Cholera Toxin Penetration and Induced Phase Transformations in Lipid Membranes. Biophys. J. 2008, 95, 629−640. (12) Krasilnikov, O. V.; Muratkhodjaev, J. N.; Voronov, S. E.; Yezepchuk, Y. V. The Ionic Channels Formed by Cholera Toxin in Planar Bilayer Lipid Membranes Are Entirely Attributable to its BSubunit. Biochim. Biophys. Acta 1991, 1067, 166−170. (13) McCann, J. A.; Mertx, J. A.; Czworkowski, J.; Picking, W. D. The Conformational Changes in Cholera Toxin B Subunit-Gangloside GM1 Complexes are Elicited by Environmental pH and Evoke Changes in Membrane Structure. Biochemistry 1991, 36, 9169−9178. (14) Nyquist, R. M.; Eberhardt, A. S.; Silks, L. A., III; Li, Z.; Yang, X.; Swanson, B. I. Characterization of Self-Assembled Monolayers for Biosensor Applications. Langmuir 2000, 16, 1793−1800. (15) Sinclair, H. R.; Kemp., F.; de Slegte, J.; Gibson, G. R.; Rastall, R. A. Carbohydrate-Based Anti-Adhesive Inhibition of Vibro Cholerae



SUMMARY AND CONCLUSIONS The data presented in this article suggests that the binding of cholera toxin to the GM1 residue appears to stabilize the structure of the membrane. The IR results show a slight decrease in the number of gauche conformers and increase in the overall tilt angle of the acyl chains when cholera toxin binds to the surface of the membrane. The major difference in the membrane structure in the absence and presence of bound CTB was a notable decrease in the relative hydration of the glycerol moiety because the massive protein layer isolates the membrane from the aqueous electrolyte. The IR measurements have also shown that the opening and closing of the CTB pore does in fact occur via the α-helical components in the center of the pore and that this pore opening is voltage-dependent. This is a significant result because it may explain the mechanism of the transport of the toxin through the membrane.



AUTHOR INFORMATION

ASSOCIATED CONTENT

* Supporting Information S

Details of experimental procedures and methods. Table 1 contains a summary of the conditions used to obtain the PMIRRAS spectral regions corresponding to various functional groups of DMPC and the cholera toxin B subunit bound DMPC/Chol/GM1 bilayers. Figures SI1−SI4 correspond to the optical constants for various functional groups of the phospholipid bilayer and the amide I band of cholera toxin B subunit. Figure SI5 shows the spectra of the nonspecifically bound CTB. Figure SI6 shows the deconvolution of the C−D stretching region at each electrode potential. Figures SI7−SI9 correspond to the Fourier self-deconvolution and derivative analysis methods for determining the number of bands associated within the complex IR regions of the bilayer and 974

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

Toxin Binding to GM1-OS Immobilized into Artificial Planar Lipid Membrane. Carbohydr. Res. 2004, 344, 1968−1974. (16) Nullmeier, M.; Koliwer-Brandl, H.; Kelm, S.; Brand, I. Interaction of Siglec Protein with Glycolipids in a Lipid Bilayer Deposited on a Gold Electrode Surface. J. Electroanal. Chem. 2010, 649, 177−188. (17) Nullmeier, M.; Koliwer-Brandl, H.; Kelm, S.; Zagel, P.; Koch, K.-W.; Brand, I. Impact of Strong and Weak Lipid−Protein Interactions on the Structure of a Lipid Bilayer on a Gold Electrode Surface. ChemPhysChem. 2011, 12, 1066−1079. (18) Zawisza, I.; Bin, X.; Lipkowski, J. Potential-Driven Structural Changes in Langmuir-Blodgett DMPC Bilayers Determined by in Situ Spectroelectrochemical PM IRRAS. Langmuir 2007, 23, 5180−5194. (19) Breimer, M. E.; Hansson, G. C.; Karlsson, K.-A.; Larson, G.; Leffler, H. Glycosphingolipid Composition of Epithelial Cells Isolated along the Villus Axis of Small Intestine of a Single Human Individual. Glycobiology 2012, 12, 1721−1730. (20) Schuck, S.; Simons, K. Polarized Sorting in Epithelial Cells: Raft Clustering and the Biogenesis of the Apical Membrane. J. Cell Sci. 2004, 117, 5955−5964. (21) Spiegel, S.; Blumenthal, R.; Fishman, P. H.; Handler, J. S. Gangliosides Do Not Move from Apical to Basolateral Plasma Membrane in Cultured Epithelial Cells. Biochim. Biophys. Acta 1985, 821, 310−318. (22) Torok, Z.; Demel, R. A.; Leenhouts, J. M. de Kruijff, B. Presequence-Mediated Intermembrane Contact Formation and Lipid Flow. A Model Membrane Study. Biochemistry 1994, 33, 5589−5594. (23) Zawisza, I.; Bin, X.; Lipkowski, J. Spectroelectrochemical Studies of Bilayers of Phospholipids in Gel and Liquid State on Au(111) Electrode Surface. Bioelectrochemistry 2004, 63, 137−147. (24) Richer, J.; Lipkowski, J. Measurement of Physical Adsorption of Organic Species at Solid Electrodes. J. Electrochem. Soc. 1986, 133, 121−128. (25) Zamlynny, V.; Zawisza, I.; Lipkowski, J. PM FTIRRAS Studies of Potential Controlled Transformations of a Monolayer and a Bilayer of 4-Pentadecyl Pyridine, a Model Surfactant, Adsorbed on a Au(111) Electrode Surface. Langmuir 2003, 19, 132−145. (26) Green, M. J.; Barner, B. J.; Corn, R. M. Real-time Sampling Electronics for Double Modulation Experiments with Fourier Transform Infrared Spectrometers. Rev. Sci. Instrum. 1991, 62, 1426−1430. (27) Barner, B. J.; Green, M. J.; Saez, E. I.; Corn, R. M. Polarization Modulation Fourier Transform Infrared Reflectance Measurements of Thin Films and Monolayers at Metal Surfaces Utilizing Real-Time Sampling Electronics. Anal. Chem. 1991, 63, 55−60. (28) Buffeteau, T.; Desbat, B.; Blaudez, D.; Turlet, J. M. Calibration Procedure to Derive IRRAS Spectra from PM-IRRAS Spectra. Appl. Spectrosc. 2000, 54, 1646−1650. (29) Bin, X.; Lipkowski, J. Electrochemical and PM-IRRAS Studies of the Effect of Cholesterol on the Properties of the Headgroup Region of a DMPC Bilayer Supported at a Au(111) Electrode. J. Phys. Chem. B 2006, 110, 26430−26441. (30) Merrit, E. A.; Sarfaty, S.; van der Akker, F.; L’Hoir, C.; Martial, J. A.; Hol, W. G. J. Crystal Structure of Cholera Toxin B-Pentamer Bound to Receptor GM1 Pentasaccharide. Protein Sci. 1994, 3, 166− 175. (31) Zhang, R. G.; Westbrook, M. L.; Westbrook, E. M.; Scott, D. L.; Otwinowski, Z.; Maulik, P. R.; Reed, R. A.; Shipley, G. G. The 2.4Ǻ Crystal Structure of Cholera Toxin B Subunit Pentamer: Choleragenoid. J. Mol. Biol. 1995, 251, 550−562. (32) Hess, B.; Kutzner, C.; Van Der Spoel, D.; Lindahl, E. GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. J. Chem. Theory Comput. 2008, 4, 435. (33) Burgess, I.; Li, M.; Horsewell, S.; Szymanski, G.; Lipkowski, J.; Majewski, J.; Satija, S. Electric Field-Driven Transformations of a Supported Model Biological Membrane - an Electrochemical and Neutron Reflectivity Study. Biophys. J. 2004, 86, 1763−1776.

(34) Spangler, B. D.; Westbrook, E. M. Crystallization of Isoelectrically Homogeneous Cholera Toxin. Biochemistry 1989, 28, 1333− 1340. (35) Kabir, S. Charge Heterogeneity of Cholera Toxin and its Subunits. FEMS Microbiol. Lett. 1986, 37, 155−162. (36) Brosseau, C. L.; Leitch, J.; Bin, X.; Chen, M.; Roscoe, S. G.; Lipkowski, J. Electrochemical and PM-IRRAS a Glycolipid-Containing Biomimetic Membrane Prepared Using Langmuir-Blodgett/LangmuirSchaefer Deposition. Langmuir. 2008, 24, 13058−13067. (37) Lee, D. C.; Durrani, A. A.; Chapman, D. A Difference Infrared Spectroscopy of Gramicidin-A, Alamethicin and Bacteriorhodopisin in Perdeuterated Dimyristol-Phosphatidylcholine. Biochim. Biophys. Acta 1984, 766, 49−56. (38) Allara, D. L.; Swalen, J. D. An Infrared Reflection Spectroscopy Study of Oriented Cadmium Arachidate Monolayer Films on Evaporated Silver. J. Phys. Chem. 1982, 86, 2700−2704. (39) Umemura, J.; Kamata, T.; Kawai, T.; Takenaka, T. Quantitative Evaluation of Molecular Orientation in Thin Langmuir-Blodgett Films by FT-IR Transmission and Reflection Spectroscopy. J. Phys. Chem. 1990, 94, 62−67. (40) Fringeli, U. P. The Structure of Lipids and Proteins Studied by Attenuated Total Reflection (ATR) Infrared Spectroscopy. II. Oriented Layers of a Homologous Series: Phosphatidylethanolamine to Phosphatidylcholine. Z. Naturforsch. 1977, 32C, 20−45. (41) Fringeli, U. P.; Gü nthard, H. H. Infrared Membrane Spectroscopy. In Membrane Spectroscopy; Grell, E., Ed.; SpringerVerlag: New York, 1981; pp 270−332. (42) Bin, X.; Zawisza, I.; Goddard, J. D.; Lipkowski, J. Electrochemical and PM-IRRAS Studies of the Effect of the Static Electric Field on the Structure of the DMPC Bilayer Supported at a Au(111) Electrode Surface. Langmuir 2005, 21, 330−347. (43) Aussenac, F.; Laguerre, M.; Schmitter, J.-M.; Dufourc, E. J. Detailed Structure and Dynamics of Bicelle Phospholipids Using Selectively Deuterated and Perdeuterated Labels. H-2 NMR and Molecular Mechanics Study. Langmuir 2003, 19, 10468−10479. (44) Blume, A.; Hübner, W.; Messner, G. Fourier Transform Infrared Spectroscopy of 13C:O Labeled Phospholipids Hydrogen Bonding to Carbonyl Groups. Biochemistry 1988, 27, 8239−8249. (45) Hübner, W.; Mantsch, H. H. Orientation of Specifically 13C O Labeled Phosphatidylcholine Multilayers from Polarized attenuated Total Reflection FT-IR Spectroscopy. Biophys. J. 1991, 59, 1261− 1272. (46) Susi, H.; Byler, D. Resolution-Enhanced Fourier Transform Infrared Spectroscopy of Enzymes. Methods Enzymol. 1986, 130, 290− 311. (47) Barth, A.; Zscherp, C. What Vibrations Tell Us About Proteins. Q. Rev. Biophys. 2002, 35, 369−430. (48) Goormaghtigh, E.; Ruysschaert, J. M.; Raussens, V. Evaluation of the Information Content in Infrared Spectra for Protein Secondary Determination. Biophys. J. 2006, 90, 2946−2957. (49) Miura, Y.; Kimura, S.; Imanishi, Y.; Umemura, J. Formation of Oriented Helical Peptide Layers on a Gold Surface Due to the SelfAssembling Properties of Peptides. Langmuir 1998, 14, 6935−6940. (50) Tamm, L.; Tatulian, S. A. Infrared Spectroscopy of Proteins and Peptides in Lipid Bilayers. Q. Rev. Biophys. 1997, 30, 365−420. (51) Worley, C. G.; Linton, R. W.; Samulski, E. T. Electric-FieldEnhanced Self-Assembly of α-Helical Polypeptides. Langmuir 1995, 11, 3805−3810. (52) Gremlich, H. U.; Fringeli, U. P.; Schwyzer, R. Conformational Changes of Adrenocorticotropin Peptides upon Interaction with Lipid Membranes Revealed by Infrared Attenuated Total Reflection Spectroscopy. Biochemistry 1983, 22, 4257−4264. (53) Okamura, E.; Umemura, J.; Takenaka, T. Molecular Orientation in Thin Langmuir−Blodgett Films of Dipalmitoylphosphatidylcholine as Studied by FTIR Transmission and Reflection−Absorption Spectroscopy. Can. J. Chem. 1991, 69, 1691−1694. (54) Marsh, D.; Müller, M.; Schmitt, F.-J. Orientation of the Infrared Transition Moments for an α-Helix. Biophys. J. 2000, 78, 2499−2510. 975

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976

Langmuir

Article

(55) Surewicz, W.; Leddy, J.; Mantsch, H. Structure, Stability and Receptor Interaction of Cholera-Toxin as Studied by FourierTransform Infrared Spectroscopy. Biochemistry. 1990, 29, 8106−8111. (56) Noda, I.; Dowrey, A. E.; Marcott, C.; Story, G. M.; Ozaki, Y. Generalized Two-Dimensional Correlation Spectroscopy: Theory and Applications. Appl. Spectrosc. 2000, 54, 236A−248A. (57) Kong, J. Y. Fourier Transform Infrared Spectroscopic Analysis of Protein Secondary Structures. Acta Biochim Biophys Sin. 2007, 39, 549−559. (58) Buffeteau, T.; Le Calvaz, E.; Desbat, B.; Pelletier, I.; Pezolet, M. Quantitative Orientation of α-Helical Polypeptides by Attenuated Total Reflection Infrared Spectroscopy. J. Phys. Chem. B. 2001, 105, 1464−1471. (59) Thompson, T. E.; Allietta, M.; Brown, R. E.; Johnson, M. L.; Tillack, T. W. Organization of Ganglioside GM1 in Phosphatidylcholine Bilayers. Biochim. Biophys. Acta 1985, 817, 229−237. (60) Mou, J.; Yang, J.; Schao, Z. Atomic Force Microscopy of Cholera Toxin B Oligomer Bound to Bilayers of Biologically Relevant Lipids. J. Mol. Biol. 1995, 248, 507−512.

976

dx.doi.org/10.1021/la304939k | Langmuir 2013, 29, 965−976